1. RAID Redundant Array of Inexpensive Disks Presented by Greg Briggs

2. RAID example: Striping Write every other block to different disk – Disk block size varies, could be 1 kB Performance – Throughput: up to N times better – Latency: same – Reliability: dropped to 1/N – Redundancy: none 123456 1 2 3 4 5 6 Disk 1 Disk 2 RAID

3. Redundancy Reliability increase instead of decrease MTTF: Mean Time to Failure MTTR: Mean Time to Repair If MTTF >> MTTR, only need to survive a small number of failures at most MTTR improved via hot spares 123456 Disk 1 Disk 2 RAID 12345 123456 6

4. RAID levels RAID levelMin # faults Example data disks Example check disks Corporations producing products -Linear / spanned080Linux, Windows 0Striped080widely used 1Mirrored188EMC, Compaq, IBM, Linux, Windows 2Memory-style ECC184none 3Bit-interleaved parity181Storage Concepts 4Block-interleaved parity181Network Appliance, Linux 5Block-interleaved distributed parity 181widely used 6P+Q redundancy282Linux (beta) 1+0 or 10 Striped mirrors188Linux 0+1Mirrored stripes188JetStor

5. RAID 1: Mirroring Optionally, let disk heads be out of sync to improve read seek time and allow parallel small reads, at the cost of write seek time Expensive (2N disks) RAID 1+0: create RAID 1 pairs, and then stripe across them RAID 0+1: create two RAID 0 sets, and maintain them as mirrors 123456 Disk 1 Disk 2 RAID 12345 123456 6

6. RAID 3: Bit-Interleaved Parity Protection group: group of N+1 disks redundantly holding particular data Inexpensive Writes go to all disks in protection group The extra disk holds parity bits, the sum (xor) of the other disks RAID 1 mirroring is like RAID 3 with N=1 123456 Disk 1 Disk 2 RAID 1&2 Parity 1&2 3&4 Parity 3&4 5&6 Parity 5&6 Disk 3

7. RAID 4 and 5: Block-interleaved parity Same # disks as RAID 3: N+1 RAID 4: parity disk is still a write bottleneck RAID 5: distributed parity removes bottleneck 123456 Disk 1 Disk 2 RAID 1 2 Parity 1&2 3 4 Parity 3&4 5 6 Parity 5&6 Disk 3 123456 Disk 1 Disk 2 RAID 1 2 Parity 1&2 3 4 Parity 3&4 5 6 Parity 5&6 Disk 3

8. RAID 4 and 5: Block-interleaved parity Read shortcut- can do small independent reads Write shortcut- just check old block and old parity

9. RAID 6: P+Q redundancy Like RAID 5, except uses N+2 disks per protection group, and can survive two disk failures 123456 Disk 1 Disk 2 RAID 1 2 Parity 1&2 3 4 Parity 3&4 5 6 Parity 5&6 Disk 3 Disk 4 Parity 1&3 Parity 4&6 Parity 5&8 78 Parity 2&7 Parity 7&8 7 8

10. Software RAID RAID normally implemented by special disk- controller cards – Transparent to OS – Hardware designed to operate with failed disks Windows and Linux implement RAID in software as well – CPU load – System may freeze with failed disks – Often, better management of RAID array – Support per-partition RAID schemes

11. RAID Summary Higher throughput, higher transaction rate Fault tolerance Thus, RAID dominates todays large storage systems